Big Data in Maize Breeding Research: Something Old, Something New, Something Borrowed.

Virtually 100% of maize cultivar development in the U.S. is done by the private sector. This situation poses existentialist questions for a professor in maize breeding and genetics. One solution around this dilemma is to use big data in maize breeding research. These big data are often borrowed and often old, but sometimes need to be augmented by newer big data found elsewhere. In this talk, we will present examples from the 1990s, 2000s, and 2010s of how different sets of big data have allowed us to answer questions about predicting hybrid performance, finding marker-trait associations, developing genomewide prediction models, and assessing genotype-environment interaction in maize. We will present both opportunities and challenges we have encountered in our use of big data in maize breeding research. We have sometimes found that carefully selected small data, rather than big data, are the answer. Overall, access to big data has affected the types of research questions that we can study and the means by which we can study such questions.